GLM-5.3 Flash / RoCE v2 / TP2, 128 GB per node / Q4_K

User reports GLM-5.3 Flash Q4_K at 10.1 t/s generation and 103.4 t/s prompt processing on two AMD Strix Halo nodes, each with 128 GB installed memory. Setup is a custom DS4 fork's staged GLM TP=2 ROCm path with a 4,608-token context allocation, 4,096-token prompt and 300-token decode at prefill batch 256, no speculative draft or MTP. The 4K point uses a cross-disciplinary prompt and is one observation; the source waived repeated timing, so no spread or confidence interval is established.